STAC Summit, 18 May 2023, Chicago
STAC Summits bring together CTOs and other industry leaders responsible for solution architecture, infrastructure engineering, application development, machine learning/deep learning engineering, data engineering, and operational intelligence to discuss important technical challenges in trading and investment.
Click on the session titles to view the slides.
|STAC Update: Machine learning|
Peter and Bishop will discuss the latest research and STAC Benchmark Council activities related to machine learning, including updates on inference and training workloads. Peter will also present results from recent STAC-ML audits.
|Is everyone a developer now? The impact of LLMs on software development|
GPT-4 and other generative AI could be creating a watershed moment for the software development life cycle. In everything from Python to C++ and even VHDL/FPGA, large language models (LLMs) are starting to help with code generation, reviews, documentation, and testing—including creating data sets to probe the edges of algorithms. No one wants to be left behind, but adoption requires answering critical questions. Will LLMs provide great leaps in productivity or only incremental assistance? Can they turn English into an effective programming language (even enabling someone with no coding experience to develop software)? Do we risk automating the errors and poor code too often found on Stack Overflow and developer blogs? Can we benefit without leaking our trade secrets or violating someone else's copyrights? Our panel of experts will examine the extent to which financial firms and their engineers can leverage LLMs to boost productivity, improve software quality, and accelerate time to market. Along the way, they'll consider security and bug risks, as well as the guardrails needed to keep our software—and markets—safe. Join them to discuss how to benefit from the bleeding edge. To kick off, some of the panelists provided a short presentation:
|"Adopting an Effective LLM Platform: Key Considerations for Success"
Alex Gonzalez, Machine Learning/Deep Learning Solutions Engineer, Hewlett Packard Enterprise
|"Large language Models for Systematic Algo/Discretionary trading incl. Alternative Data"
Prabhu Ramamoorthy, CFA, FRM, CAIA, Customer/Partner Developer Relationship Manager, NVIDIA
|STAC Update: Big workloads|
Bishop and Peter will present the latest STAC Benchmark Council activities in storage and compute critical workloads, including STAC-M3 (the full stack needed for enterprise tick analytics) and STAC-A3 (strategy backtesting).
|Can object stores lift your data?|
Trading and investment firms must gain an advantage over the vast and ever-growing data they process, store, and analyze. Object stores, originally used for low-cost snapshots and archives, now bring space and cost savings to mission-critical applications. Could they be the lever that financial firms need? For object stores to provide leverage, everyone from storage architects to software developers needs to understand them. Our panel of experts will discuss important aspects, including use cases the performance is suitable for, how get/pull differs from read/write, integrating existing storage solutions and analytic stacks, and how moving this cloud-native technology on-prem impacts hybrid systems. Join us to explore how object stores can help you shoulder the weight of data. To kick off, some of the panelists provided a short presentation:
|"OBJ at the speed of light!"
Michael Rothenberg, Global UDS Business Development, Financial Trading & Energy, Dell Technologies
|"Architectures for Explosive Data Growth - Efficiently Handling AI, Analytics and Distributed Data"
Keith Miller, VP Technical Sales, Services and Support, DDN
|"Pure Innovation: A new Era of Scale"
Rob Glanzman, Global Strategic Alliances Principal Architect, Financial Services, Pure Storage
|STAC Update: Cloud Communications|
Peter will discuss the latest STAC Benchmark Council updates on defining cloud communication benchmarks for on-prem-to-cloud and cloud-to-cloud communications
|Trading ecosystems in the cloud: Architecting for the inevitable?|
As we've discussed at past STAC events, exchanges and central service providers are moving important functions into the cloud. Matching in traditional asset classes may take a while to migrate, but the logic for many surrounding functions is compelling. Trading firms will either want or need to follow suit, which means architecting for cloud has become critical. This raises many questions for an industry that has relied mostly on on-premises solutions. Which functions are likely to become most common in the cloud, over what time horizon? What kind of throughput, latency, and consistency can the cloud provide? How will the technologies that financial firms have come to rely on be adapted to the cloud? How should architects handle infrastructure decisions such as whether to go hybrid cloud or multi-cloud? Join our panelists as they dive into these questions and yours.
|Is the cloud ready for low-latency messaging?|
Low and consistent messaging latency are fundamental requirements for trading infrastructure. Kernel bypass technologies and UDP—both often discussed at STAC events—are dominant tools in on-prem, low-latency messaging but are largely absent in public clouds. Can trading workloads migrate to the cloud without them and still be successful? James thinks success is possible if you embrace different techniques and leave the "lift-and-shift" approach behind. He'll examine common issues when communicating between processes and containers over a software-defined IP network. He'll also cover emerging patterns that help overcome the challenges, including kernel bypass in the cloud, IPC, and simulated multicast. Along the way, he'll address how developers can modify software to take advantage of these patterns and achieve fanout. Be sure to bring your questions for James as he explores if the cloud can support your messaging needs.
|Ethernet multicast congestion control: What works and what does not.|
Network engineers building reliable multicast on high-speed ethernet depend on hardware and software mechanisms to avoid packet drops and latency spikes, but this dependency isn't always well-founded. In fact, while porting their infrastructure from Infiniband to ethernet, Deutsche Boerse engineers have encountered several issues with multicast congestion control. They have found that while congestion control often works for unicast, the same cannot be said for multicast. For example, the back pressure Deutsche Boerse has relied on for years to slow down senders and avoid retransmissions during microbursts is implemented differently from vendor to vendor, if implemented at all. Christoph will discuss the issues they've uncovered and his experience working with different hardware, protocols, and vendors to develop a solution. Bring your questions for Christoph and learn from his experience.
|STAC Update: Fast data|
Peter will discuss updates from the FPGA special interest group and the latest results and STAC Benchmark Council activities related to low-latency/high-throughput realtime workloads.
|Can trading transition to 25G?|
Migration to 25G, 40G, and 100G Ethernet is in full swing among enterprises and cloud providers, but exchange connections and trading infrastructures have resolutely remained at or below a 10G ceiling. It is tempting to believe this will always be so. But ever-increasing data rates and the need for larger data sets create mounting pressure for more bandwidth. Will trading be able to resist the pressure forever? Dave will explore whether the industry can change to 25G, along with several related questions. What impact will 25G have on low-latency infrastructure, not to mention timestamping, data capture, and operational support? Will firms use forward error correction, given the latency hit, and if not, how can they manage the data risk? How will more bandwidth impact market microstructure, and will it enable more insight into exchange behaviors? Join Dave as he looks at a possible future of trading networks.
|FPGAs beyond ultra-low latency?|
It's no secret that FPGAs are a favorite component for ultra-low latency system engineers. Firms can gain a significant edge by placing relatively simple trading strategies in an FPGA to avoid user space entirely. But what role can FPGAs play in more complex, compute-hungry strategies? Can FPGAs combine with other processors in quant-based trades or critical pricing algorithms? For that matter, does FPGA have a home outside real-time decision-making, such as in research infrastructures? Join our panelists as they discuss these questions and yours on the future of FPGAs in capital markets. To kick off, some of the panelists provided a short presentation:
|"FPGA Hardware Acceleration in Electronic Trading, AI, and Data Analytics"
Matt Certosimo, Data Center FPGA Field Application Engineer, AMD
|"The Edge of the Edge: PCB Latency in Trading Reaction Time"
Adam Sherer, Verification Technology Executive, Cadence Design Systems
If you have speaker or topic suggestions, please submit the form below:
About STAC Events & Meetings
STAC events bring together CTOs and other industry leaders responsible for solution architecture, infrastructure engineering, application development, machine learning/deep learning engineering, data engineering, and operational intelligence to discuss important technical challenges in trading and investment.
Satya PandeyWilliam Blair Jeff CaldwellARB TG
Satya PandeyWilliam Blair
Jeff CaldwellARB TG
Dr. Christoph LameterDeutsche Boerse David LariviereUniversity of Illinois
Dr. Christoph LameterDeutsche Boerse
David LariviereUniversity of Illinois
Zachary RichardsDRW Michael RothenbergDell Technologies
Michael RothenbergDell Technologies
Matt CertosimoAMD Alex GonzalezHewlett Packard Enterprise
Alex GonzalezHewlett Packard Enterprise
Keith MillerDDN Alastair RichardsonAMD
Rob GlanzmanPure Storage Kevin FormbyKeysight Technologies
Rob GlanzmanPure Storage
Kevin FormbyKeysight Technologies
Dave SnowdonArista Networks Adam ShererCadence Design Systems
Dave SnowdonArista Networks
Adam ShererCadence Design Systems
Mike McLaughlinVast Data Josh ChouRefinitiv, an LSEG Business
Mike McLaughlinVast Data
Josh ChouRefinitiv, an LSEG Business
James WatsonAdaptive Prabhu RamamoorthyNVIDIA
Paula ClarkeKX Cliff MaddoxNovaSparks
Tom StephensTechary Laurent de BarryExegy
Laurent de BarryExegy
Vahan SardaryanLDA Technologies Jean-François GagnonOrthogone
Vahan SardaryanLDA Technologies